In a surprising development that has shaken the semiconductor industry, DeepSeek’s creators have unveiled a large language model system that cost less than $6 million—an amount far below the investments made by U.S. tech giants like OpenAI and Meta Platforms. Despite these modest expenses, DeepSeek’s AI system has shown performance benchmarks that rival or surpass those of leading chatbots. DeepSeek’s AI Assistant rapidly gained popularity, quickly becoming the most downloaded free app in the U.S. on the Apple App Store, surpassing OpenAI’s ChatGPT.
This breakthrough sent shockwaves through the industry, with Nvidia, a prominent AI chipmaker, suffering a dramatic 17% drop in shares, losing $600 billion in market value following DeepSeek’s debut. Industry analysts have characterized the launch as a wake-up call for all players in the AI space. While some details about DeepSeek’s R1 model remain unclear and potentially exaggerated, experts believe this marks the beginning of a new competitive arms race within the AI sector.
Despite the immediate panic, Morgan Stanley analysts believe there is no cause for alarm, seeing the selloff as a buying opportunity. The firm has maintained its “Top Pick” rating for Nvidia, with a price target of $152. While acknowledging potential long-term risks associated with DeepSeek’s rise, Morgan Stanley points to strong near-term prospects, particularly with Nvidia’s Hopper and Blackwell chips.
Meanwhile, Citi has lowered its price target for Nvidia from $175 to $163 but maintains a buy rating ahead of the company’s earnings report due February 26. Citi projects in-line results for the January and April quarters, though it warns that Nvidia’s gross margins may see a decline.
The debate surrounding DeepSeek intensified when U.S. lawmakers proposed legislation that would ban the chatbot from government-owned devices due to national security concerns. The bipartisan bill, led by Representatives Darin LaHood and Josh Gottheimer, follows an analysis revealing that DeepSeek’s code could potentially send user information to China Mobile, a state-owned telecommunications company. Several U.S. federal agencies, including the Navy and NASA, have already restricted the app, and Texas became the first state to ban it from government devices.
Internationally, concerns have spread beyond the U.S. Governments in South Korea, Australia, and Italy have taken steps to limit DeepSeek’s use, citing security risks and the possibility of sensitive data being exposed to foreign governments. Australia’s Department of Home Affairs recently banned the app, citing unacceptable security risks posed by the app’s data collection practices.
Despite the controversies, analysts have expressed skepticism about DeepSeek’s claims of low-cost innovation, noting that similar advancements are likely being explored by top AI research labs worldwide. Penn Mutual Asset Management has emphasized that, despite DeepSeek’s arrival, U.S. tech companies continue to invest heavily in AI infrastructure, with Microsoft, Meta, Alphabet, and Amazon already spending billions in 2024 alone. AI-related expenditures are expected to exceed $250 billion in 2025, fueling competition and innovation in the sector.
In the rapidly evolving AI race, experts agree that DeepSeek’s emergence serves as a stark reminder for big tech firms to maintain their focus on efficiency and innovation or risk being overtaken.
